Contrastación de los resultados con estudios similares

Overall, this work shows that each of the dimensions are important to legislators in providing representation to constituents. However, little is known about the circumstances that change the relative importance of each dimension on either the supply or demand sides. Past work has not typically incorporated all four in the same model. This is a problem because legislators with limited time and energy are not able to provide all four types of representation at a maximum level, but instead must make choices between them. As Fenno notes: “[b]uilding a reelection constituency at home. . . requires time and energy. Inevitably, these are resources that might otherwise be [given to policy-related] efforts” (1978, 215, see also Hogan N.d.). Studies of only one dimension miss this crucial complexity inherent in the practice of representation.

In the chapters that follow I address this problem by developing a more comprehensive theoretical model accounting for citizen preferencesandlegislators’ choices between these four dimensions. I then derive several hypotheses based on this unified theory and test them with data from survey experiments administered to citizens, survey experiments ad- ministered to American state legislators, and data from state legislators’ websites. Overall, I demonstrate that beyond simply structuring policy preferences, expectations on the role of government influence citizen preferences for how legislators should provide representa- tion. Further, I show that the choice of how to prioritize the dimensions of representation reflects strategic considerations by legislators.

3 A SUPPLY AND DEMAND THEORY OF REPRESENTATION

3.1 Introduction

Eulau and Karps (1977) were some of the first to make the point that scholars too often conceptualize representation exclusively as responsiveness to policy. Since the time of their work literature has developed on the four dimensions of the concept that are the focus of this dissertation.1

• Policy: Responding to district policy concerns through position-taking, bill introduc-

tions, or voting behavior in the legislature.

• Service: Assisting constituents who need help with government agencies.

• Allocation: Securing government funding for projects in the district.

• Descriptive: A connection through shared identity traits like gender or race.

However, as is shown in chapter 2, extant literature has typically focused narrowly on only one of these dimensions rather than looking comprehensively at all of them. The result is several relatively separate literatures rather than a single, integrated body of research on representation. The need for work connecting these different dimensions is apparent from two different observations.

The first is that legislators see their jobs as being comprised of multiple elements. The interviews I described in chapter 1 provide anecdotal evidence consistent with this asser- tion. More generally, it is also supported by the fact that American legislators typically win re-election despite several factors that obstruct the basic policy congruence relation- ship. These obstacles include low levels of constituent knowledge and awareness (Griffin

1_{I focus on dyadic (individual) rather than collective representation (e.g., Weissberg 1978) because service}

and Flavin 2007), interest groups and bureaucracy (Lowi 1979), pressure from the party or legislators’ own preferences (Mayhew 1974), institutional factors (Lax and Phillips 2012), variance in electoral competition (Fiorina 1974; Griffin 2006), or constituency traits like opinion heterogeneity (Bailey and Brady 1998). To compensate for these obstacles, leg- islators likely emphasize different types of representation to develop trust in the district (Fenno 1978).2

Moreover, focusing narrowly on one dimension of representation assumes, either im- plicitly or explicitly, that constituents want, and thus legislators provide, the type of repre- sentation that is the topic of study. For instance, research on policy-based representation assumes constituents are satisfied only by policy congruence. Figure 3.1 shows that this assumption is problematic. The graph plots the distribution of responses to a question I included on the 2010 Cooperative Congressional Election Study (CCES) about three jobs of a state legislator. Specifically, I asked respondents to rank the importance of attention to policy, service, and allocation.3 _{Notice that none of the three response options is strongly}

preferred over the others; the proportions range only from 29.9% to 36.4%. This shows that there is considerable variation in what people think legislators should be doing to provide representation.

In sum, many studies on only one dimension omit important elements of the process (i.e., the other dimensions). Doing so is at odds with the fact that legislators view the job as being comprised of many elements and fails to account for variation in constituent demand for representation. In the remainder of this chapter I develop a theoretical framework to

2_{Of course, this does not mean that policy behavior is irrelevant. Indeed, scholars show that legislators’}

issue choices can have real electoral consequences (Canes-Wrone, Brady, and Cogan 2002; Hogan 2008). Nonetheless, the relative safeness of incumbent legislators suggests that many of those who are “out-of-step” still gain support through other means.

3_{The question wording was as follows: “Here is a list of some activities that occupy political representa-}

tives as part of their job. We want to know how important you think these activities are for [state legislators]. Please rank these activities in order of importance. (1) Learning about constituents’ opinions in order to better represent their views. (2) Helping constituents who have personal problems with government agencies. (3) Making sure the district gets its fair share of government money and projects.

Figure 3.1: Rankings of Three Jobs of a State Legislator by a Sample of American Adults

1stChoice 2ndChoice 3rdChoice

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Propor tion Policy Service Allocation 95% Confidence Interval

Note: The graph plots the distribution of responses to a question on the 2010 Cooperative Congressional

Election Study (CCES) about three jobs of a state legislator (N = 977). Specifically, I asked respondents to

rank the importance of attention to policy, service, and allocation.

explain both variation in citizen preferences for and in legislator supply of the dimensions of representation.